Optimization Algorithms Of Cross-layer Congestion Control For Wireless Sensor Networks

Wireless sensor networks (WSNs) integrate the sensor, micro-electromechanical system, modern computing and wireless communication networks. WSNs have attracted much focus from both academic and industrial fields due to their great application prospect. However, sensor nodes in WSNs are severely constrained by the limited computing power, limited storage capacity and limited communication capabilities. Therefore, energy efficient is the most important issue of concern in WSNs. In this dissertation, three key technologies of WSNs are studied, including congestion control, routing protocols and scheduling. The main purpose is to optimize network structure, save the network energy, improve energy load balancing of nodes in the network, prolong the network lifetime, improve network security and ensure the reliability of the network.(1) Cross-layer congestion control algorithms based on sliding mode variable structure and data driven theories in control theory are presented, to solve the problem of congestion caused by limited processing ability and channel competition in WSNs.Min-data package of node output traffic is preferentially transmitted in link-level congestion; and active queue management method is used for congestion control of node-level. The controller can adaptively adjust the data transmission rate according to local congestion condition of network nodes, and adaptively allocate MAC channel. The validity of the proposed algorithm is shown by the Lyapunov function.Data driven theory is applied for aiming at difficult to obtain linear model in WSNs. Firstly, linearized models based on I/O are obtained with pseudo-partial-derivatives in WSNs. Secondly, as to the linearized models, new cross-layer congestion control algorithms are studied:min-data package of node output traffic get higher priority to possess the channel, if a channel is occupied by several nodes simultaneously; while the others select rest channels. The controllers implement the purpose that congestion control can be simultaneously carried in Media Access Control (MAC) layer and transport layer. It can adaptively adjust the data transmission rate according to local congestion condition of network nodes, and reasonably allocate MAC channel. The algorithm dramatically restrains the congestion in WSNs, and significantly maintains higher throughput and lower delay, also effectively improves the quality of service for the whole network.(2) A cross-layer congestion control algorithm based on compressed sensing is proposed. Cross-layer congestion control for WSNs based on compressed sensing (CS) is considered for which node-level congestion and link-level congestion are occurred simultaneously. The first group constructs a WSNs control model. The second group includes controller designing: sparse signals are mapped compressively in bottleneck node, which reduces the data transmission in bottleneck nodes; then the compressed signals are reconstructed with convex optimization theory, whose target is reconstruction of the original signal. Via Lyapunov function the validity of the proposed algorithm is shown. There are several improvements in the packet loss rate and throughput as compared to the other congestion control protocols. The proposed scheme achieves congestion control in transport layer and MAC layer, also improves the quality of service for the whole network.(3) A cross-layer optimal design based on compressed sensing is proposed for congestion in WSNs. The design objective is achieving a cross-layer optimal design for signal input, routing selection and energy allocation. The reasonable optimization model is decomposed into three subsections for three layers in WSNs:congestion control in transport layer, scheduling in MAC layer and routing algorithm in network layer, respectively. The three functions mutually interact and are regulated by congestion ratio so as to achieve a global optimality. Congestion control can be robust and stable by compressed sensing (CS) theory. Routing selection is abided by fair resource allocation principle. The resources can be allocated more and more to the channel in the case of not causing more severe congestion, which can avoid conservatively reducing resources allocation for eliminating congestion. This method guarantees the higher throughput, the more accurate ratio of CS.(4) Considering data compressed rapidity and fair channel allocation in WSNs, a fast cross-layer optimal algorithm based on compressed sensing technique is proposed. Considering congestion problem by robust optimization with congestion ratio for two classic aspects in energy limited WSNs:minimizing the bit transmission and maximizing the information transmission. To achieve the goal, three protocols are developed. The first protocol, the desire of control input, is designed based on compressed sensing technique. A minimal bit of signal is provided to reduce the transmission flow for congestion model. The second protocol is resource allocation. The resources can be allocated more and more to the channel in the case of not causing more severe congestion, which can avoid conservatively reducing resources allocation for eliminating congestion. Channel selection is abided by fair resource allocation principle. The above protocols are impacted through a congestion ratio at network layer, transport layer and MAC layer, respectively. In addition, sensing matrix is constructed by structurally random matrices method before data transmission, in which signal could be compressed very fast, decreasing data communication, and curtailing the time of data compression and reconstruction deeply. Secondly, the improved Homotopy method is applied for signal reconstruction in Sink node, in which the high accuracy reconstruction signal is obtained only through kth iteration. The proposed algorithm effectively restrains congestion, and achieves higher throughput and lower energy consumption, the congestion time is greatly cut down et al.(5) Considering energy constraint in WSNs, cross-layer energy efficient optimal algorithms are proposed based on compressed sensing. A packet error rate control strategy based on congestion rate is designed, which makes the packet error rate in the domain of validity, and can be used power "to each according to his needs". Second, the optimal schemes of control input and link capacity with compressed sensing are discussed. Third, we construct minimum energy problem with power control function, congestion control function, link allocation function and so on. These functions interact through and are regulated by congestion rate so as to achieve a global optimality. Then, using a Lagrange optimization model, in which by adjusting power, rate, channel selection and channel capacity functions to optimize the aforementioned model, and we can accommodate services as well as exploit the tradeoff between efficiency and fairness of resource allocation. The proposed algorithms achieve maximization of network performance in relieving congestion and saving energy.Lastly, the summary of the whole dissertation is given and the research directions in future are put forward.